Weekend Long Read: Five Ways AI Will Put China Ahead

At a latest summit, I gave a speech that reviewed the event of synthetic intelligence (AI) over the previous years and predicted that within the coming 20 years, the mixture of AI and different new applied sciences would result in 5 main industrial modifications that can permit China to steer the world in science and expertise.

The modifications are as follows:

1. China will lead the world in superior manufacturing as AI empowers automation upgrades

2. China’s provide chain will dominate the world as vitality and materials costs plummet

3. Sensible cities and the Web of All the pieces (IoE) can be realized and totally computerized driving will turn out to be universally accessible

4. As China’s improvements in enterprise intelligence multiply, AI will allow a brand new order of enterprise operations

5. AI mixed with well being care improvements will decrease mortality from illness and prolong individuals’s lives

Judging by the point and assets wanted to provide new applied sciences — from exploration to software to commercialization — I might say that there’s a very excessive chance of those modifications occurring in China within the subsequent 20 years or so.

The next is an edited model of the speech.

Catching up with the U.S.

By way of realizing industrial worth, three years in the past I predicted that China’s AI would meet up with the U.S. round 2023. I primarily based this on plenty of factors:

·To begin with, China has a outstanding expertise benefit, with scientists like academician Yao Qizhi (姚期智), professor Zhou Zhihua （周志华） and chief scientist of Sinovation Ventures Zhou Ming （周明）, in addition to their sensible college students

·Second are the enterprise mannequin and knowledge benefits extracted from China’s large market

·Third are sturdy and pragmatic authorities insurance policies supporting the AI business

·Fourth, China’s aggressive startup ecosystem has given rise to sturdy, progressive native firms whose rapidly-growing companies are driving equally speedy advances in knowledge, AI and different applied sciences.

Trying again on developments over the previous three years, we are able to already see that these predictions are coming true. There are 4 waves of AI presently at work reshaping each business.

·The primary wave is web intelligence. Pushed by AI, web giants comparable to Meituan and ByteDance have created miracles pushed by their large market and super-high market worth.

·The second wave is enterprise intelligence. The applying of AI in finance, insurance coverage and different fields with an enormous quantity of taggable knowledge has achieved outstanding outcomes.

·The third wave is intelligence within the bodily world. The recognition of laptop imaginative and prescient, listening to and numerous IoT-connected units is empowering AI notion of the actual world to rival that of people — in some methods it even exceeds our talents. Though purposes presently make use of fairly weak AI, they’ve nonetheless surpassed people’ capability in sure duties comparable to facial recognition, merchandise recognition, video viewing and clever high quality management.

One of many hottest areas today is the fourth wave — totally automated intelligence. When AI has fingers and ft, together with imaginative and prescient and notion, it may well generate appreciable worth in eventualities comparable to unmanned driving, robotic arm gripping and drone spraying.

After all, regardless of the general growth of the AI business, concerted efforts are wanted to resolve a few of China’s arduous expertise bottlenecks.

Though China could lack related Nobel Prize and Turing Award winners, it has been, when it comes to the variety of high-quality papers on AI and different key areas, shortly catching up lately, approaching and even surpassing the extent of the U.S. or Europe in some areas. Such advances make us very optimistic concerning the prospects for bettering China’s total analysis capability.

Prediction 1: China will lead the world in superior manufacturing as AI empowers automation upgrades

Up to now 12 months, regardless of the pandemic and commerce friction with the U.S., China efficiently maintained its place because the world’s manufacturing unit. Nonetheless, we should admit that as Chinese language labor is already twice as costly as that of India and Vietnam, its low-cost benefit is progressively disappearing. Over time, these different nations will purchase know-how and presumably come to problem China’s standing.

Easy methods to drive manufacturing upgrades whereas defending labor earnings?

Industrial automation and intelligence will do wonders for China’s manufacturing business. What’s presently achieved by human fingers, ft, eyes and brains will progressively get replaced with full automation. Automation can convey down working prices whereas permitting China to keep up its benefits in manufacturing high quality, course of and provide chain. Chinese language factories presently have the world’s largest variety of industrial robots, however their intelligence degree remains to be fairly low. Sooner or later, robots can be progressively upgraded with AI and broadly utilized in quite a lot of eventualities.

The businesses invested in by Sinovation Ventures usually are not essentially attempting to interchange employees. Having served as an unbiased director at Foxconn for a few years, I perceive that it’s very tough to interchange the employees who make iPhones. That’s as a result of guide dexterity, hand-eye coordination and the delicate annual iteration of the iPhone make it very tough to implement an automatic method. However I’ll be daring sufficient to foretell that within the following industries, we’ll determine some eventualities notably conducive to automation even utilizing right this moment’s expertise.

AI has already surpassed the human eye within the area of business imaginative and prescient. Deep studying is developed with convolutional neural networks, so it’s able to carrying out facial and object recognition in addition to clever high quality inspection — one of many “lowest hanging fruits” of AI empowerment.

Agriculture is one other “lowest hanging fruit” for AI empowerment. Drones can sow seeds, apply fertilizer and water fields with unmanned tractors. Whether or not a farm is rising cotton, rice or wheat, it’s simpler to realize standardization there than it’s for manufacturing cell telephones, footwear and so forth. As soon as profitable implementation is achieved, AI plus agriculture can thus develop globally with a lot larger ease.

Autonomous driving has accelerated in the wake of the Covid-19 pandemic, however Stage 5 totally automated driving stays hardly ever seen. However, in restricted eventualities comparable to an unmanned manufacturing unit or warehouse dealing with, or the place human employees collaborate with robotic arms, worth could be created instantly.

Within the life sciences, MegaLab — one other investee of Sinovation Ventures — has efficiently utilized AI plus automation. Within the context of Covid-19, for instance, its clever automation was utilized to nucleic acid testing in hospitals, facilities for illness management and laboratories, together with full automation of the post-sampling testing course of.

Almost 1,000 robotic arms work at a Chongqing automobile plant in January 2019.Picture: VCG

Horizontally, business would be the first space to see breakthroughs. China will be capable to preserve its necessary place because the world’s manufacturing unit by making use of “automation plus intelligence” upgrades for enterprise and residential.

It’s simple for related expertise to unfold from business into the industrial sector. If expertise is mature sufficient to automate the motion of an industrial forklift, then we’re additionally able to making a mall safety robotic or a home housekeeping robotic.

Similar to human eyes and fingers, robotics should additionally undergo a strategy of growth from easy to advanced. What we should do first is direct capital to resolve the business’s hardest issues. A better AI, having been skilled by a considerable amount of knowledge and undergone mass manufacturing, will enter the enterprise area and industrial eventualities that includes end-user interplay, then lastly enter properties, with totally different types of by-product merchandise.

Vertically, laptop imaginative and prescient is a comparatively simple matter, adopted by comparatively crude handling and moving things, and at last — essentially the most tough step — the simulation of fingers. Progress can be made in all of those areas over the subsequent 20 years. By that point, I consider robots will be capable to self-replicate, self-repair and be suitable with quite a lot of 3D printers. The period of future robotics will start with the upgrading of China’s automated factories earlier than penetrating into extra automated enterprise and residential eventualities.

Prediction 2: China’s provide chain will dominate the world as vitality and materials costs plummet

The second prediction is concerning the vitality revolution.

The price of photovoltaic (PV) energy technology has fallen to five% of what it was 20 years in the past, whereas the price of lithium batteries has fallen to round 2%. At this time, PV is already one of many most cost-effective technique of producing electrical energy. With the growth of new-energy autos, PV energy will solely get cheaper. Photo voltaic vitality absorbed however not used is wasted. As vitality storage expertise can be advancing, vitality could be saved in lithium batteries for later use.

On this means, electrical energy will turn out to be extra energy-saving, environmentally-friendly and simple to retailer. Photo voltaic and lithium battery applied sciences will reduce the worth of electrical energy by 1 / 4 to a 3rd each 10 years. At this fee, in 20 years, the price of electrical energy will solely be a tenth what it’s right this moment.

When the day comes for a change in the way in which that vitality is produced, which nation can have the most effective benefit?

Up to now, nations with oil, coal mines and pure fuel reserves had a pure benefit. However right this moment, as photo voltaic or wind plus lithium battery applied sciences mature, and given the abundance of photo voltaic or wind vitality on the planet, it will likely be nations with the strongest manufacturing capabilities and industrial capacities that take the lead.

Progress in new supplies will push the vitality revolution but additional, whereas artificial biology and lots of different applied sciences will convey the price of manufacturing considerably down. Within the agricultural sector, the figuring out issue will shift from cultivation and harvesting capability to produce manufacturing capability. Nations with sturdy manufacturing will thus paved the way in agriculture as properly.

An agricultural professional inspects vegetation on the Langkun Agricultural Science and Know-how Park in Binhai, Jiangsu province, in December 2020. Picture: IC Picture

Sooner or later, vertical farming can be a key “breadbasket” for human subsistence. Greens and fruit can be grown in factories, soil-free, on vertically stacked planters. Rising applied sciences comparable to plant-based meats and stem cells can create the meals of the longer term with out harming animal life.

Within the foreseeable future, the supplies and vitality required to provide a brand new merchandise will turn out to be cheaper and greener, manufacturing can be accomplished by totally automated robots and mankind will enter an period of abundance, by no means to see poverty or starvation once more. Reaching this period of abundance, nevertheless, would require manufacturing power on the world-factory degree. And whether or not or not China can lead the world to such an period relies on the diploma and pace of automation and clever upgrades, with AI, massive knowledge and different platform-level applied sciences as the brand new underlying infrastructure.

Prediction 3: Sensible cities and the Web of All the pieces can be realized and totally computerized driving turn out to be universally accessible 

The third prediction is the belief of sensible cities and the Web of All the pieces, together with the popularization of totally automated driving.

I divide the period of totally automated driving into two phases:

The primary stage is its realization in easy, particular eventualities to pave the way in which for commercialization. For instance, we are able to begin with computerized forklifts in warehouses, then fixed-route shuttle buses and minibuses, adopted by taxis. When these eventualities are realized, we can be within the Stage 4 stage of autonomous driving.

The following stage is Stage 5, totally automated driving. On this stage, automated driving can change people in all site visitors eventualities and circumstances.

This present day could are available in simply 20 years. How can we speed up the method? The secret’s constructing sensible transportation and sensible cities.

Not too long ago, a brand-new expertise has been utilized to some new freeway initiatives in Jiangsu province: at nighttime or poor climate circumstances, computerized floor lights and sensors assist cost new-energy autos. These are the sorts of strategies that may assist pace the arrival of sensible transportation and sensible cities.

Relating to autonomous driving, attitudes in Europe, the U.S. and China diverge fairly sharply. In Europe and the U.S., persons are merely enthusiastic about find out how to get self-driving automobiles on the street whereas leaving the roads and cities themselves unchanged. China, in the meantime, has a powerful openness to the brand new expertise revolution in addition to highly effective modes of execution. Folks listed below are questioning why the roads and the cities should stay the identical. Why not let the freeway “discuss” to the automobile? Why not layer automobiles and pedestrians from beneath to above floor in new city facilities, completely avoiding vehicle-pedestrian collisions and accelerating the arrival of Stage 5 autonomous driving?

As soon as the period of unmanned driving is right here, it will likely be matched with an epoch-making working system (OS). How did the previous few generations of OS come about? The Home windows OS emerged with the applying of PC laptop software program, whereas iOS and Android OS emerged with the speedy world popularization of smartphones after 2015. As soon as an OS takes off, it spreads to succeed in every nook of the world. What, then, will immediate the beginning of the subsequent nice working system? I feel it’s more likely to be totally automated driving. That’s as a result of, so as to obtain totally automated driving, we’d like to have the ability to combine AI, a number of sensing applied sciences, {hardware} and software program communication and operational choices in a really highly effective computing setting combining native, distant and cloud computing.

Such an OS is extremely tough to comprehend, however as soon as it has been, our autos can have “eyes, fingers and ft” and be capable to “hear, stroll and transfer” like an individual.

As soon as autonomous driving reaches this degree, two issues will observe. First, huge quantities of knowledge can be opened up, permitting for higher-function AI calculations; second, the dimensions of automated automobile manufacturing will develop, thus bringing down the price of required parts. The worth of LIDAR, for instance, will drop considerably in 20 years. Finally, the price of constructing a wise automobile will come down, together with the worth of constructing a robotic, and the decreased prices can be conducive to autonomous driving’s full-scale popularization. Within the not-so-distant future, it is possible for you to to function a self-driving plus robotic OS at an affordable worth.

Prediction 4: As China’s improvements in enterprise intelligence multiply, AI will allow a brand new order of enterprise operations

The fourth prediction is an progressive breakthrough in enterprise intelligence (BI) in China. Future BI is concerning the understanding of human language — not simply language recognition, however the precise understanding of textual content’s content material.

There are 4 layers of knowledge intelligence: on the backside is fragmented and unstructured knowledge, and above that, an higher layer of knowledge that “reads” the info content material. Most deep studying right this moment makes use of these “knowledge and data” layers.

The third layer, “data,” is presently utilized in mapping and relational chains, however most of that is but to be utilized to AI. The very best layer is “perception,” the place clever evaluation and assisted decision-making primarily based on occasions and insights can provide customers suggestions. Nonetheless, this fourth layer is much from good.

We consider that the way forward for BI lies within the prime layers — that’s, in “data and perception.”

“Personalization” is one in all AI’s strengths. At this time, for instance, quick video app Douyin offers “customized companies” by continually recommending movies to people primarily based on their previous viewing habits. Taobao, the e-commerce website, in the meantime, is ready to continually advocate merchandise that match customers’ preferences. However why is AI “personalization” just for current movies, and never new ones? If producing new movies is simply too tough, we are able to begin by attempting to create “customized” textual content. For instance, if the person likes cute Japanese types, upon understanding their feelings, wants and use eventualities, AI can work as a magnificence advertising advisor, making suggestions for the person particularly.

One other state of affairs is “monetary occasion understanding.” At this time’s finance is extremely digitalized, with danger management and cash laundering measures, however is it attainable to take issues to the subsequent degree? For instance, what ought to we spend money on after we know for certain that the worldwide pandemic recedes in about two months? What shares ought to I promote if I do know the Panama Canal is blocked for one more three weeks? These are the questions main monetary buyers are asking every single day. Options enhanced by search engines like google and yahoo, together with “data and perception” refined by AI and massive knowledge, have the potential to forge the longer term’s new enterprise order. Subsequently, the large knowledge and AI revolutions, that are solely getting began, must be formatted and intellectualized to supply true “insights.”

Prediction 5: AI mixed with well being care improvements will decrease mortality from illness and prolong individuals’s lives

At current, the life sciences are present process large modifications worldwide and medical data is quickly being digitized. Along with the recognition of wearable units, sure medical processes comparable to AI viewing, gene sequencing and different new applied sciences are bringing construction and standardization to huge quantities of latest knowledge. Knowledge, the required gas for AI growth, will definitely convey higher and extra plentiful alternatives for progressive medical purposes of AI.

Medical doctors give a affected person a distant ultrasound examination in Huzhou, East China’s Zhejiang province in October 2019. Picture: VCG

For instance, right this moment we often solely discuss to a health care provider for 5 minutes after we go to the hospital, however there’s a large quantity of knowledge hidden in these 5 minutes. If a health care provider, with the assistance of AI, may gather each element, they may make a “customized” prognosis and therapy plan to optimize the method, scale back prices and enhance effectivity. 

Along with the usage of AI and massive knowledge to supply “customized” therapy, we’ll see massive alternatives for the common software of surgical robots and use of AI to invent new medicine.

Sooner or later, AI could convey two large modifications to prescribed drugs: the primary is a significant discount within the prices and time concerned in pharmaceutical manufacturing. It presently prices $2 billion to make a medication, however sooner or later, the price is predicted to drop to a tenth of that determine. These days, excessive prices are making pharmaceutical firms reluctant to develop medicine for uncommon illnesses. However the aforementioned AI-supported drug discovery course of from Insilico Medication dramatically reduces the time it takes to develop conventional medicine within the pre-clinical stage — from 4 to 5 years to 18 months — which in flip drastically reduces time and value investments concerned in drug discovery! Sooner or later, AI can considerably speed up the event of latest medicine and relieve extra sufferers’ illnesses.

The second change is “customized” therapy plans. Why does everybody use the identical medication once they get sick? Utilizing conventional pharmaceutical strategies, it prices $1 billion to $2 billion to fabricate a drug, from the R&D to the scientific stage. Sooner or later, we can make a drug with simply $100 million to $200 million. When that occurs, extra medicine can be out there to assist customized therapeutic therapies. Seeing how China has embraced and invested in new well being care applied sciences, the nation is predicted to steer the economic transformation of AI and well being care.

Conclusion 

If the U.S. may turn out to be a worldwide hegemon by succeeding in main in 4 key sectors of the economic revolution — vitality, manufacturing, transportation and well being care — a century in the past, then right this moment, China is well-poised to keep up its place because the “world’s manufacturing unit” by advantage of its benefits in massive knowledge, AI, automation and clever manufacturing capability. As talked about above, manufacturing benefits are what drive vitality revolutions and modifications in agricultural provide; thus, as soon as manufacturing turns into stronger, vitality, agriculture and future new supplies are likewise strengthened in flip. At this time, AI and well being care is on a quick observe for growth, whereas new applied sciences comparable to artificial biology will current actual alternatives.

These 5 predictions could also be solely a part of the story. However there is no such thing as a doubt that China, the “world’s manufacturing unit,” stands to be a significant participant within the subsequent 20 years of worldwide industrial change and generate monumental worth for the entire world.

This commentary was translated from the writer’s weblog. 

Kai-fu Lee is chairman and CEO of Sinovation Ventures.

Contact editor Joshua Dummer (joshuadummer@caixin.com)

Download our app to obtain breaking information alerts and browse the information on the go.

Get our weekly free Should-Learn publication.